Application of ANN model to predict the sediment removal efficiency of silt extractor

The excess sediment load in canals generally causes the silting problem. The vortex chamber type silt excluders are used to extract the excess sediments from the canals to increase its discharge carrying capacity. The silt removing efficiency of silt excluder is generally calculated using the available empirical equations. The prediction results of the existing empirical equations were not found satisfactory. To improve the prediction results, a new model was developed using the artificial neural network (ANN). The experimental data were trained to developed the ANN model. The performance of developed ANN model was evaluated using two different statistical parameters MSE and R. These performance parameters suggest that the accuracy of developed ANN model improved significantly than that of other available empirical equations. The R value was found to be equal to 0.9863 which is quite closer to 1. In addition to this, the removal efficiency predicted using the ANN model lies within the error band of +/- 20% which is more accurate than that predicted by existing empirical relationships in which the estimated removal efficiency lies within the error band on +/- 40%.